<AWT Dev> Ideas about improved architecture of the generic Swing peers

[Roman Kennke]

Thanks Oleg, I really like scepticism :-)

> Based on our previous experience with text components in XAWT? I'd say that
> this can be hard (very hard or even impossible( to write GOOD AWT components
> using Swing.  By "good" I mean components which will compatible with current
> behavior of the same component.

I see this is a problem, but nothing that can't be solved with some
hacking :-)

>   Also I wonder if AWT component will
> have  AWT LaF
> when it is being used in application which uses both AWT and Swing.

Eh, I was not aware there was something like AWT LaF. I thought AWT just
uses whatever the backend uses (Motif, Win32, or in the case of GNU
Classpath, GTK or Qt). Having the same LaF for Swing and AWT should be
good IMO. But if not, it should be possible to install a special UI
class on the backing Swing components...

> So, I'd suggest to run JCK tests on this new toolkit and be a little
> bit more skeptic and
> try to find weak places of this approach.

Surely I am. Regarding the JCK - yeah funny. Give me access to it, and
I'll do it :-)

>   Because if you want to
> incorporate this toolkit to
> JDK you must have compatible behavior, LaF, and pass JCK.

I can solve the first two points (compatibility and LaF) but not the
last one :-)

Thanks for giving me some perspective.

Cheers, Roman

-- 
http://kennke.org/blog/
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: Dies ist ein digital signierter Nachrichtenteil
Url : http://mail.openjdk.java.net/pipermail/caciocavallo-dev/attachments/20080906/bf3d71af/attachment.bin